WP2. Epistemic Risk, Epistemic Luck, and Cognitive Achievement

Our second work package will focus on the epistemological implications of the preceding one. Ernest Sosa’s dominant reliabilist virtue epistemology (2007; 2015; ms.) employs the concepts of luck and risk quite recurrently without offering an explicit definition of them. We believe that Sosa's view could be reinforced by properly elucidating those concepts along the lines described in the previous work package. In particular, we will defend the hypothesis that ‘animal knowledge’ is justified true belief that is not affected by veritic luck, whereas ‘reflective knowledge’ is justified true belief where veritic risk is below some standards. The main goal of this second work package is thus to defend this reading of Sosa’s virtue reliabilism, and to show its strengths over its main competitor: Pritchard’s anti-risk virtue epistemology (Pritchard 2016).

In order to elaborate and defend this hypothesis, we will hinge on our previous defenses of virtue epistemology (Broncano-Berrocal 2014; 2017; 2018; forthcoming-a; Navarro 2015). In particular, we will defend the view that luck and risk accomplish importantly distinct roles in epistemic evaluation because they have different social functions in our social assessments of achievements, epistemic or otherwise. In a nutshell: the qualification of achievements as ‘lucky’ implies a responsibility- diminishing claim (either by reducing or by modulating the merit or fault of the agent in the performance), whereas by assessing achievements as 'risky’ we are essentially tracking the lack of safety of the performances in question, a matter that is orthogonal to the agent’s merit or fault in such a lack (i.e., sometimes risk is a motivation for praise and compliment, whereas luck could hardly be so). We aim to show that this different social role of risk attributions can be applied to the twin aims of attaining truth and avoiding error, in that it allows distinguishing between the first and the second-order level. Nevertheless, we will also discuss the possibility of accounting for avoidance of error by means of what we will call precautionary virtues—virtues whose function is to stop belief formation when the circumstances are such that errors are likely—which might perhaps not require any second-order evaluation of sorts.

Finally, WP2 will deal with some literature that focuses on the relationship between luck, risk and achievements—see Bradford 2015 for monograph on the notion of achievement. We will defend the existence of internal conceptual links between luck attributions, agent’s intentions and the lack of causal deviant chains in the actual occurrence of performances. We will introduce our account in contrast to extant views in the literature that lack such an agential perspective, such as McKinnon’s statistical account of luck (2013; 2014) or Pritchard’s purely modal account of luck and risk (2015; 2016). The view that we hold defends that there is a variety of risk assessment that is tightly connected to the possibility of human error, and that is inadequately grasped by any account that dodges the role of intentional action. Our hypothesis in this respect is that a twofold argument may help motivate this point: on the one hand, we will resort to the notion of virtue as it is understood in virtue reliabilism and argue that the peculiar kind of failure not grasped by views that omit appealing to intentional action takes place in the kind of conditions where an agent is expected to manifest aptness; second, we will use a version of the so -called ‘generality problem’ for reliabilism against such views. By means of these two ideas, we aim to defend that there are essentially indexical reasons for preferring being the one ‘in charge’ of one's performance—i.e., motivations that are good reasons for me, in so far as I am the agent doing the action, reasons that I do not have when I am just a bystander observer.